DESCRIPTION: Barrett's esophagus is a metaplastic condition that develops in an estimated 10-20% of persons with long-standing gastroesophageal reflux disease (GERD). Patients with this condition are at high risk, estimated at 1-2% per year, of developing esophageal adenocarcinoma, a rapidly fatal cancer that, for unknown reasons, has risen sharply in incidence since the mid-1970s. Although much research has been directed toward identifying predictors of progression i patients with Barrett's esophagus, there is little information about the causes of the conditions itself. The first Specific Aim of this case-control study is to compare cases with newly diagnosed Barrett's esophagus (N=335) to controls from the general population (N=335) to test the hypothesis that specific environmental exposures and host factors increase risk of Barrett's metaplasia, and to estimate the fraction of cases attributable to them. Also to be examined will be the role of obesity, diet high in fat and low in fruits and vegetables, low serum vitamin C, vitamin E, carotenoids and selenium, tobacco use, alcohol consumption, use of medications that promote reflux and family history. The second Specific Aim is to identify determinant of Barrett's esophagus among individuals with severe and persistent GERD, by comparing cases to patients undergoing upper endoscopy for reflux symptoms, but who are biopsy-proven negative for Barrett's esophagus (N=335). Specifically, the hypothesis to be tested is that among patients with GERD: 1) cigarette smoking, alcohol consumption and dietary intake of nitrosamines are associated with increased risk of Barrett's esophagus, and that higher dietary intake and serum levels of antioxidants are associated with decreased risk; and 2) duodenogastric reflux (measured as the concentration of bile in fasting gastric juice) is associated with an increased risk of metaplasia. The prevalence of short-segment Barrett's will be estimated among patients undergoing endoscopy for GERD, using the largest sample of endoscoped patients, to date, all biopsied according to a standard protocol. Given that 95% of persons with Barrett's esophagus remain undiagnosed and that esophageal adenocarcinoma survival rates are dismal, the most direct means of reducing esophageal adenocarcinoma incidence amy be by preventing metaplasia, rather than by the surveillance and treatment of patients already diagnosed with the condition.